- Definition of different central tendencies
- Real life example
Talking about central tendencies, in which cases the median and the mode can be more useful than the mean? Illustrate through real life examples.
?Definition of different central tendencies
The notion of central tendencies represents the average of the set of values.
It is a basis calculation when you have to analyze a series of values. However, you have to pay attention because there exists three types of central tendencies. In each case, the result is different because the calculation is not exactly the same.
The average that everybody knows is the mean. It is the result of the sum of the values divided by the number of values (generally named N).
The median is determined by the central value of the range. The way to find it is different depending on the number of total values. In fact, if the value' number of the series is odd, then the median of the series will be corresponding to the middle value. If the value' number of the series is even, then the median of the series will be corresponding to the mean of the two middle value of the series.
The mode is defined by the value that happens more frequently in a series of data.
[...] As we said previously more we have data more forecast are accurate. Nobody knows future. But there are difference between forecasting future (i.e. tomorrow) in a short-term period and forecasting future (i.e. in 50 years) in a long-term period. Data we used to forecast an event on the week will be more accurate and precise than data we used to forecast an event in 50 years. It is why the longer the horizon, the less accurate the forecast. For example, forecasts for the French president election in 2012 are more accurate than forecasts for election in 2017 or 2049 because we don't know how will be the French population and expectations at this moment and we don't know who wants be president? [...]
[...] The method is simple. It is consists in multiplying forecast by a seasonal factor. The main method steps are: 1. Determine the sample mean of past data Divide each past data by the sample mean (obtained in the step 1). We obtain normalized values (with respect to sample mean) Determine seasonal factors for each period by doing the mean of normalized values for each period Calculate de-seasonalized data applying linear Finally, multiplying seasonal factor with de-seasonalized data (calculated just before) in order to determine the forecast for a given future period. [...]